Results:
Baseline patient characteristics are given in Table 1 . Patients with DMS were older than RMS patients (77 ± 12 vs. 61 ± 15 y, p < 0.001). Study subjects were predominantly females (80% in DMS vs. 83% in RMS, p = 0.7). Patients with DMS had a higher prevalence of hypertension (92% vs. 70% in RMS, p = 0.009) and chronic kidney disease (53% vs. 29% in RMS, p = 0.045). There was a higher prevalence of AF in the RMS group (38% vs. 14% in DMS, p = 0.034). Patients with DMS had a lower mean TMPG (6.0±1.8 vs. 7.9±2.9, p<0.003) and a larger MVACEQ (1.4±0.4 vs. 0.98±0.3 cm2) despite a similar LVOT SV (in the absence of significant MR and aortic regurgitation) (Table 2 ). After the exclusion of patients with ESRD, there were 16 patients in the DMS group with high LVOT flow defined as LVOT VTI >25 cm. Even after the exclusion of high LVOT flow, the DMS group still had a larger MVACEQ (1.3±0.4 vs. 1.0±0.2 cm2, p<0.001) and lower TMPG (5.6±2 vs. 8±3 mm HG, p=0.001) compared to RMS group.
The average of mean TMPG in all subjects was 6.3 mmHg. Patients with a TMPG of > 6.3 mmHg (n=39), had a larger MVACEQ in the DMS group (n=24) than RMS group (n=14) (1.4±0.4 vs. 0.8±0.2 cm2, p= 0.001) (Figure 3 ). The frequency of low flow state defined as SV index <35 ml/m2 was 41% (n=24) and 43% (n=10) in the DMS and RMS groups, respectively (p=0.8). TMPG was significantly lower in the DMS group irrespective of the SV index (Table 2 ). We stratified the echocardiographic variables by the presence of AF in the DMS group. No significant in-between group differences (DMS patients with AF versus without AF) were observed in mean TMPG (mean difference -0.50 mmHg [-1.9 - 0.90], p = 0.475), LVOT SV (mean difference -7.2 ml [-17.1 - 2.0], p = 0.122), and DMSI (mean difference 0.02 [-0.06 - 0.87], p = 0.705).
We stratified DMS patients by MVACEQ to understand the association of MS severity with other echocardiographic variables (Table 3 ). There was a statistically significant gradual increase in E velocity with worsening MS severity defined by MVACEQ. However, A velocity and E/A ratio did not correlate with MS severity. There was a statistically significant difference in DMSI between all three groups, and it progressively got smaller with severe stenosis.
DMSI accounted for 79% of the variation in MVACEQ(Table 4, Figure 4A ), whereas the mean TMPG accounted for only 8% of the variation in patients with DMS (Figure 4B ). There was also no correlation between DMSI and mean TMPG in the DMS group (Figure 4C ).
The ROC curve showed that a DMSI of ≤ 0.50 was associated with MVACEQ ≤ 1.5 cm2 (AUROC 0.971, p = 0.001) and DMSI of ≤ 0.351 was associated with MVACEQ ≤ 1.0 cm2 (AUROC 0.965, p = 0.0001).
The interclass correlation (performed in 8 patients) for mean TMPG was 0.909 (CI 0.546 – 0.982, p < 0.001), for MVACEQ was 0.966 (CI 0.828 – 0.993, p < 0.001), and for DMSI was 0.980 (CI 0.899 – 0.996, p < 0.001) suggesting excellent reliability for all measurements. The inter-reader variability for MVACEQ was 13%, mean TMPG was 14%, and DMSI was 7%.
Survival and mortality data were available for all subjects in the DMS group. 27 ±18 months of follow-up was available after the echocardiography study used for study enrollment. The all-cause mortality rate was 53% in the DMS group and 41% in the RMS group at an average survival duration of 13 ± 14 vs. 34±14 months after the enrolment echocardiography. The Cox hazard regression model revealed that the left atrial volume index and ESRD were independent predictors of mortality in patients with DMS. Kaplan-Meier curve analysis showed a non-significant trend towards worse survival when the cutoff values of MVACEQ ≤ 1.0 cm2 and DMSI ≤ 0.35 were used (Figure 5 ). Other echocardiographic parameters such as MVACEQ ≤ 1.5 cm2, mean TMPG > 7 mmHg, or SV index ≤ 0.35 did not predict mortality risk.